vaibhav upreti 0864a261c1 fix(tools): honor string true for SKILL.md disable-model-invocation ...

YAML `true` already worked; quoted `"true"` / `yes` / `1` did not, so authors who copied the docs as a string still attached the skill.

2026-08-24 20:26:14 +01:00

..

agent

refactor(agent): rename ConclusionHandler to ConclusionParser (#5375) (#5413)

2026-08-23 12:14:44 +05:30

agent_harness

refactor(agent_harness): inject the investigation runner instead of a module global (#5682)

2026-08-24 17:10:15 +01:00

domain

refactor(reporting): make any tool's output citeable via record_evidence_entry (#5518)

2026-08-23 13:41:44 +01:00

llm

feat(concurrency): make the LLM client cache thread-safe and pool investigation workers (#5685)

2026-08-24 18:20:04 +01:00

messages

…

state

…

tool

refactor(investigation): evidence mapper travels with the tool (#5514)

2026-08-23 13:06:39 +01:00

tool_framework

fix(tools): honor string true for SKILL.md disable-model-invocation

2026-08-24 20:26:14 +01:00

__init__.py

…

context_budget.py

fix(core): give GPT-5.4 the 1M window instead of the gpt-5 128k catch-all

2026-08-24 20:26:13 +01:00

events.py

…

llm_invoke_errors.py

fix(core): keep Azure deployment-not-found investigation text generic

2026-08-24 20:26:14 +01:00

provider.py

…